External Email - Use Caution
Hi FS team,
I am running a morphometry study on a public dataset with most, but not all, subjects having a T1 and T2 (about 75% have both, remaining only have T1). To get the best surfaces possible, I have been including the T2 in recon-all process when able to. However, this means subjects with only a T1 have underwent a different processing workflow. ICV is an important predictor in my study, and I was wondering if the presence of the T2 image in recon-all should impact the ICV estimates? As I see it, I have the following options:
1) T2 is a factor, so I use only include T1 in recon-all so all subjects have the same pipeline 2) T2 is a factor, so I harmonize the data as I would with data from different sites/scanners 3) T2 is a factor, so I only analyze subjects with T1 + T2 4) T2 is not a factor, just use all subjects with no special considerations
I am also running vertex-wise surface based morphometry, where I imagine pial refinements could be more impactful. I suppose in that case, I should harmonize the morphometry data, but was curious if you think something like that is necessary.
Thank you, Steven Meisler
External Email - Use Caution
No, it does not affect it. The ICV is actually and *estaimated* ICV derived from the registration of the T1 to MNI305 space, so the T2 does not come into play.
On 4/17/2023 2:11 PM, Steven Meisler wrote:
External Email - Use CautionHi FS team,
I am running a morphometry study on a public dataset with most, but not all, subjects having a T1 and T2 (about 75% have both, remaining only have T1). To get the best surfaces possible, I have been including the T2 in recon-all process when able to. However, this means subjects with only a T1 have underwent a different processing workflow. ICV is an important predictor in my study, and I was wondering if the presence of the T2 image in recon-all should impact the ICV estimates? As I see it, I have the following options:
- T2 is a factor, so I use only include T1 in recon-all so all subjects have the same pipeline
- T2 is a factor, so I harmonize the data as I would with data from different sites/scanners
- T2 is a factor, so I only analyze subjects with T1 + T2
- T2 is not a factor, just use all subjects with no special considerations
I am also running vertex-wise surface based morphometry, where I imagine pial refinements could be more impactful. I suppose in that case, I should harmonize the morphometry data, but was curious if you think something like that is necessary.
Thank you, Steven Meisler
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://secure-web.cisco.com/1iAD3fj03-Svj5Q9Ac12rWbor-MUY4-pHjCtS5cnKdh5Nny...
freesurfer@nmr.mgh.harvard.edu
-
Douglas N. Greve -
Steven Meisler